OpenAI’s ChatGPT introduces advanced voice mode

OpenAI has recently launched an advanced voice mode for ChatGPT, which has garnered positive feedback from a limited group of ChatGPT Plus subscribers. The new feature, released a few days ago, demonstrates the AI’s capabilities in singing, accent imitation, language pronunciation correction, and narrative storytelling. Various clips showcasing these abilities have appeared online.
One example features a user, @nickfloats, asking ChatGPT to tell a story in the style of an airline pilot addressing passengers. The AI responded almost immediately, adjusting its audio to mimic an intercom. While the AI encountered challenges with more complex tasks, such as incorporating engine sounds, the voice itself was clear and emotive, and the AI handled user interruptions well.
In a conversation uploaded to YouTube, ChatGPT claimed it could process inputs in “dozens of languages,” although the exact number varies depending on dialects and regional variations. In one clip, ChatGPT corrects the pronunciation of French words, offering specific tips on inflection. Another video shows the chatbot telling an emotive story in Turkish, with some native speakers noting that while the accent did not sound native, the AI was able to convey the story and react appropriately by laughing and crying at certain points.
The bot also attempts regional US accents, with one video showcasing examples from New York, Boston, Wisconsin, and a stereotypical “valley girl.” Other videos demonstrate ChatGPT’s advanced voice feature singing in different styles, including a blues-style rendition of “Happy Birthday” and imitating animal sounds.
Several different male and female voices were used in these demonstrations, though the Scarlett Johansson-like “Sky” voice, which was removed from the service in May, was not included.